A new optimization algorithm based on the integration of reinforcement learning and HHO algorithm for virtual machine placement in cloud data centers A new optimization algorithm based on the integration of

• Published in: Computing Vol. 107; no. 6
• Main Authors: Ghasemi, Arezoo, Tejani, Ghanshyam G.
• Format: Journal Article
• Language: English
• Published: Vienna Springer Vienna 01.06.2025
• Subjects: Artificial Intelligence; Computer Appl. in Administrative Data Processing; Computer Communication Networks; Computer Science; Information Systems Applications (incl.Internet); Regular Paper; Software Engineering; Cloud computing; 68T20; 68W50; 68T05; Harris–Hawks optimization; Machine learning; 90C59; 68M14; VM placement
• ISSN: 0010-485X; 1436-5057
• DOI: 10.1007/s00607-025-01493-0

Cloud computing has significantly transformed the use of computing resources, but it still faces major challenges, especially in virtual machine (VM) placement, which directly impacts the efficiency of cloud data centers. Various methods based on metaheuristic or machine learning algorithms have been proposed to address this issue. However, each approach has its strengths and weaknesses. For example, metaheuristic algorithms are prone to getting trapped in local optima, while machine learning techniques often face challenges in defining key components, which are essential for achieving effective convergence. To overcome these limitations, this paper introduces a novel method that integrates the Harris Hawks Optimizer (HHO) with reinforcement learning, combining the strengths of both approaches. In this method, the exact approach is used to replace random selection methods for choosing the appropriate mechanism to generate a new solution. The proposed method outperforms traditional techniques, achieving up to 25% better energy efficiency, reducing resource waste by 27%, and improving load balancing by 20%. Experimental results, conducted across four distinct scenarios in both homogeneous and heterogeneous data centers, demonstrate the superior performance of the proposed method compared to existing solutions.